Background In the vertebrate retina, six types of neurons and Muller glia originate from common progenitor cells. Extracellular signals and intrinsic regulatory mechanisms dictate cell-type specification in the neural retina. Our studies have yielded insights into transcriptional regulation of photoreceptor development. However, molecular pathways controlling differentiation of retinal neurons are still poorly understood. We had previously established that rod photoreceptor fate is largely determined by the basic motifleucine zipper transcription factor NRL. Thus, we have focused our efforts on the NRL-centered gene regulatory network that determines rod photoreceptor development. We are exploring three complementary aspects: (i) Targets of Nrl that dictate rod identity and function, (ii) The temporal delay in maturation of rods, and (iii) Upstream regulators of Nrl expression. We are using a combination of in vitro and in vivo approaches and high throughput methods. Transcriptional regulation of photoreceptor development Among the genes significantly down-regulated in Nrl-/- retina, we had identified receptor accessory protein 6 (Reep6) as a direct Nrl target and confirmed this observation by chromatin immunoprecipitation and sequencing (ChIP-Seq). We identified a novel isoform of Reep6 in the retina (Reep6.1), using exon specific Taqman assay and rapid analysis of cDNA ends (5-RACE). The Reep6.1 protein includes 27 additional residues encoded by a novel exon 5 and is specifically expressed in developing and mature rod photoreceptors. By ChIP, we identified NRL binding within Reep6 intron 1. Reporter assays in cultured cells and transfections in retina explants revealed an intronic enhancer sequence that mediates Nrl-directed Reep6 expression. Based on functional studies in mouse and zebrafish, we implicated Reep6 as a key downstream molecule in Nrl-centered control of rod photoreceptor differentiation and homeostasis. Rod spherules and cone pedicles exhibit distinct morphology and functional connectivity with bipolar neurons. We observed that synaptic terminals in Nrl-/- retina switch from spherule to pedicle-like morphology, with reduced connections with rod bipolars. We hypothesized that NRL controls expression of genes involved in presynaptic assembly and synapse formation in developing mouse retina. We have selected 20+ candidate genes relevant to synaptic development in rod photoreceptors, based on differential expression between Nrl-/- and wild type retina, NRL ChIP-Seq data, and developmental expression pattern. By in vivo electroporation of shRNA, we have narrowed our list to 6 genes that appear to modulate rod photoreceptor synapse formation. Mouse Prickle genes, Pk1 and Pk2, are differentially regulated by NRL, suggesting the hypothesis that they have a role in the mammalian retina where they could establish planar cell polarity (PCP) and contribute to normal development of axons and dendrites. In the mature retina, Pk1 is highly expressed in cholinergic amacrine neurons. Pk1 knockdown in neonatal retina leads to defects in inner and outer segments and axon terminals of photoreceptors supporting our working hypothesis (2). Stringent control of rhodopsin expression is critical for photoreceptor maintenance and survival. While NRL and CRX bind to the proximal promoter, the rhodopsin enhancer region (RER) may contribute to high and precise expression of rhodopsin. We identified RER-bound proteins by mass-spectrometry. We validated the binding of NonO (p54nrb), that is implicated in coupling transcription to splicing, and of three NonO-interacting proteins hnRNPM, Ywhaz and Ppp1ca. NonO and its interactors can activate rhodopsin promoter and function synergistically with NRL and CRX. DNA binding domain of NonO is critical for rhodopsin promoter activation. ChIP-Seq showed high occupancy of NonO at rhodopsin and a subset of rod phototransduction genes, suggesting an important contribution of NonO and its interacting proteins in regulating rod specific gene expression and homeostasis. Epigenetic regulation of photoreceptor development To explore the role of epigenetic marks in retinal development, we generated conditional knockout alleles of DNA methyltransferase 1 (Dnmt1) in mice. Dnmt1 knockdown in early eye development induced cell fate changes only in cones. However, photoreceptors completely lacked outer segments (OS) despite near normal expression of phototransduction and cilia genes. Retinal pigment epithelium (RPE) morphology and polarization were also disrupted as early as embryonic day (E)15.5. Defects in OS biogenesis were evident with Dnmt1 conditional knockout in RPE, but not in photoreceptors, suggesting a unique function of DNMT1-mediated DNA methylation in controlling RPE apico-basal polarity and in guiding photoreceptor OS morphogenesis. Furthermore, we detected a reduction in DNA methylation of LINE1 elements (a measure of global DNA methylation) in developing mutant RPE compared with neural retina, and of Tuba3a, which exhibited dramatically increased expression in mutant retina. Our results establish an instructive role of RPE in photoreceptor OS formation (4). The phenotype of Dnmt1 mice suggested that DNA methylation may be involved in photoreceptor development. Thus, we sought to establish genome-wide DNA methylation (meDNA) profile of rod photoreceptors by meDNA-Seq. When combined with RNA-Seq expression profiles from the same cell population at the same time points, our preliminary data suggest an inverse correlation between DNA methylation and RNA expression. We are also investigating the role of histone methylation in regulating gene expression in developing photoreceptors. We first developed a method that enables genome-wide profiling of histone signature in small numbers of purified rod photoreceptors and then applied the new method to survey changes in histone signature during rod development. Small scale ChIP-Seq for the H3K4me3 (active mark) and H3K27me3 (repressive mark) revealed a late acquisition of H3K4me3 at many phototransduction gene loci, consistent with late onset of expression. Furthermore, an initial H3K4me3 enrichment with little/no H3K27me3 at some cone gene loci followed by complete loss of H3K4me3 and high enrichment of H3K27me3 upon maturation supports the default S-cone pathway hypothesis we proposed earlier. Next Generation Sequencing data analysis and integration As we designed more experiments and generated large amount of diverse sequencing data, it became clear that we needed to invest in standardized and reliable experimental methods as well as robust analytical work flows. As a result, we have developed a new protocol for dissociating and flow-sorting of GFP +ve rods from NrlGFP mice. Furthermore, we have designed RNA-Seq analysis work flow that combines several publicly available tools, selected based on the species and the dataset under study. We believe we have reached a stage where we can start integrating data from RNA-Seq, transcription factor ChIP-Seq, histone methylation ChIP-Seq, and meDNA-Seq to address questions on gene regulatory networks in mouse photoreceptor and retinal development. Finally, we have initiated the collection of similar datasets from human embryonic stem cells (hESC)- and induced pluripotent stem cell (hiPSC)-derived retinal organoids to extend our question to human retinal development. Significance Our studies provide new insights into photoreceptor biology and disease. We are defining the role of specific genes and epigenetic signals in establishing photoreceptor identity and function during retinal development. This project complements our efforts to engineer stem cells into photoreceptors and functionally integrate them into the disrupted architecture of degenerating retinas after transplantation (EY000474).